=== JanC is now known as Guest39941 | ||
=== JanC_ is now known as JanC | ||
=== maclin1 is now known as maclin | ||
mterry | Saviq: is there a way we could have CI run on MPs that are WIP? (maybe with an opt-in or opt-out...) I would find it helpful to have CI run the tests for me and make sure the MP is all-green before unleashing it on reviewers (I can run locally, but it's not always the same and takes some of my time to babysit them) | 13:37 |
---|---|---|
om26er | Hi! How far along are pointer friendly indicators ? Are they expected to appear anytime soon ? | 14:20 |
cimi | om26er, hi, yeah they should happen soon https://bileto.ubuntu.com/#/ticket/2291 | 14:32 |
Saviq | mterry, it's just software, wonder if it's a common enough use case to warrant the added complexity? | 14:50 |
Saviq | +but | 14:50 |
Saviq | mterry, in any case, lp:jenkins-launchpad-plugin | 14:50 |
tsdgeos | Saviq: mterry: why not doing what you did in https://code.launchpad.net/~mterry/unity8/slm-test/+merge/313546 ? | 15:02 |
tsdgeos | i seems to have worked since noone reviewed that branch :D | 15:02 |
mterry | tsdgeos: it still creates noise | 15:03 |
tsdgeos | mterry: well, then you can just run the CI "manually", no? | 15:03 |
tsdgeos | otoh that's also a bit cumbesrome | 15:03 |
mterry | tsdgeos: not as user friendly yeah | 15:03 |
mterry | This isn't a huge pain in my ass. Just trying to streamline my flow and avoid situations where reviewers (you for example) wait for me to get green lights. I guess I can run tests more locally but that's not as nice as CI doing it for me | 15:04 |
tsdgeos | sure | 15:05 |
Saviq | mterry, I think that's probably the most interesting thing - improving the unity8-ci job so that it takes just the MP link and finds all it needs itself | 15:11 |
Saviq | instead of having to pass branch and MP link and revision... | 15:11 |
Saviq | you can already pass only the branch, it will work, just not report anywhere (of course) | 15:12 |
mterry | Saviq: oh that's part of it being user unfriendly. But also a normal MP gives me an email from LP when it's done, automatically kicks off on a simple bzr push, I can see the history of the job runs in the MP, etc. | 15:13 |
Saviq | mterry, sure, lp:j-l-p is the thing that's deciding which MPs to run CI for | 15:14 |
mterry | I might poke at it, yah | 15:14 |
Saviq | more important, IMO, is so it runs on top-approved MPs than on WiPs, but I suppose this might be easier | 15:14 |
Saviq | mterry, as for "maybe opt-in or something", that would suggest LP should have a new status between WiP and NR | 15:15 |
Saviq | not sure I'd go as far as that | 15:15 |
Saviq | I'd probably just do --include-wip | 15:15 |
Saviq | it only runs for branches owned by trusted people anyway | 15:15 |
Saviq | there's not so many WiP ones | 15:16 |
Saviq | and it's just a matter of adding a "Work in Progress" string to a list | 15:16 |
Saviq | top-acked are a bit more tricky, because you want to un-ack them and make sure the top revision is what's approved | 15:16 |
mterry | Saviq: maybe -- or you could opt in by setting Description to "ci-plz" or something hacky... | 15:17 |
Saviq | oh and yeah, git ;) | 15:17 |
mterry | But all CI could work too | 15:17 |
mterry | I mean | 15:17 |
mterry | All WIP could work too | 15:17 |
Saviq | I'd be ok with that | 15:17 |
Saviq | there are not so many of those | 15:17 |
Saviq | and it'd be a per-project decision | 15:18 |
tsdgeos | we actually have kind of a lot of WiP | 15:18 |
tsdgeos | but almost noone commits to them | 15:18 |
tsdgeos | so that's "fine" | 15:18 |
mterry | If anyone complains about the extra churn, I'd also be happy if it just did WIP on mterry branches :) | 15:18 |
Saviq | tsdgeos, yeah that's what I meant, there's little movement on them | 15:20 |
Saviq | when we turn this on is when things will be busy for a while, but it will settle down | 15:21 |
Saviq | and most of them will probably die in conflicts to start with | 15:21 |
tsdgeos | yep | 15:21 |
tsdgeos | i wouldn't complain if we enable for trusted-WiP either | 15:21 |
tsdgeos | can enable it on a friday so all the busy stuff is done over the WE | 15:21 |
* tsdgeos did a bit of french there | 15:21 | |
mterry | josharenson: so I looked at the arrangement branch a bit... nothing jumped out at me as obviously wrong. I tried to back up the branch to find where it went wrong, but Mir dependencies made that hard. no insights yet | 17:07 |
josharenson | mterry: I'm in the same place.. there is only 1 small thing I noticed | 17:08 |
josharenson | mterry: I have all kinds of Keys.onPressed scattered through out the code | 17:08 |
josharenson | mterry: and there was a weird thing with the app drawer... If I typed ~3 characters in the search bar, there were hundreds of key events triggered | 17:09 |
mterry | huh | 17:10 |
josharenson | mterry: which was suspicious... but I backed out the changes to the shell and the launcher and saw no improvement | 17:10 |
josharenson | mterry: I even replaced the greeter with a blue rectangle and didn't see any change | 17:10 |
mterry | huh | 17:10 |
josharenson | mterry: Heres one for you... Log in (with a passwordless user since you can't type) and then lock the session... Greeter works perfectly afterwards. | 18:03 |
* josharenson has an idea or two now... | 18:03 | |
mterry | josharenson: that's odd | 19:31 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!